University of Groningen
Real-time positron emission tomography for range verification of particle radiotherapy
Ozoemelam, Ikechi
DOI:
10.33612/diss.133158935
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Ozoemelam, I. (2020). Real-time positron emission tomography for range verification of particle radiotherapy. University of Groningen. https://doi.org/10.33612/diss.133158935
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
129
Chapter 6
Summary and Outlook
6.1 Introduction
The accurate delivery of dose to tumours with minimal irradiation of healthy tissues is the main promise of charged particle therapy (CPT). The realization of this goal is, however, challenged by among others range uncertainties and as a result the delivered dose distribution deviates from the intended one. In vivo range verification techniques, which detect secondary emissions arising from the interaction of charged particles with the body, have been proposed as quality assurance tools to guide the accurate positioning of the Bragg peak. Positron Emission Tomography (PET) of the beam-induced positron emitters such as 15O (T1/2 = 122 s) and 11C (T1/2 = 1218 s) has been used for monitoring of therapy in clinical studies. Due to the half-lives of these nuclides, provision of prompt feedback on dose delivery is precluded. In this thesis, a study of the production of suitable very short-lived positron emitters for in vivo range verification of helium beam radiotherapy and of real-time PET imaging of these nuclides for monitoring therapy with both helium ions and protons is presented.
6.2 Real-time imaging of short-lived positron emitters during
helium beam radiotherapy
The last two decades have seen a strong increase in the use of charged particles, mostly protons and carbon ions, in radiotherapy (Jermann 2017). Recently, following previous use at Lawrence Berkeley Laboratory in the 1970s, there has been a resurgence of interest in therapy with helium ions (Durante and Paganetti 2016, Tommasino et al 2015, Kempe et al 2006, Knäusl et al 2016, Grün et al 2015). There are now plans for implementation of treatments with helium ions in centres such as the Heidelberg Ion Beam Therapy Center (HIT) (Krämer et al 2016, Mairani et al 2016 and Tessonnier et al 2018). The rationale for the renewed interest in helium ions stems from the “middle-ground” advantages helium is supposed to have over the commonly used protons and carbon ions. With such interest and considering that helium ions, like every other particle therapy ion, are sensitive to range uncertainties, it becomes relevant to study the feasibility of prompt feedback using PET-based imaging of short-lived positron emitters. The feasibility was addressed in two different investigations. Firstly, the production of short-lived positron emitters during helium beam irradiation was studied and their relevance assessed in the context of achieving prompt feedback was assessed. Then, an imaging experiment was performed to evaluate the precision in range measurement that can be obtained when imaging the positron emitters produced. The main results from these studies as presented in chapter 3 and 4 respectively of this thesis are summarized below.
6.2 Real-time imaging of short-lived positron emitters during helium beam radiotherapy
6.2.1 Production of very short-lived positron emitters for PET imaging of
helium radiotherapy
The integral yield of short-lived positron emitters was explored by irradiating thick targets of graphite (carbon), water (oxygen), phosphorus and calcium with 50 MeV/u 4He and 59 MeV/u 3He ions, which have nearly the same range in matter. Using a NaI(Tl) detector, the integral yield of relevant nuclides was obtained by detecting, where present, the unique gamma photons associated with their decay and by a half-life analysis of the time spectra of the 511 keV annihilation photons. The integral yields were corrected for the effects of positrons escaping the experimental targets as well as photon attenuation in the targets.
The production of short-lived nuclides, including 13O/12N (T1/2 = 8.6 ms/11.0 ms) on both water and graphite targets was observed with both ions. Analysis of the time spectrum of the coincidence events, obtained during the imaging experiments with 4He beam as summarized in section 6.2.2, shows the production of a very short-lived contribution with half-life of 11.1 ± 0.3 ms. The half-life of the activity contribution seen during irradiation with 4He suggests that almost exclusively 12N is produced on PMMA. The more accurate identification of the contributing nuclide is attributed to the higher efficiency of the detectors, reduced irradiation times, thus less influence of longer lived nuclides in the fitting and the use of coincidence measurements. A more conclusive statement on the nuclide produced during irradiation with 3He would require additional investigations. In proton irradiations, Dendooven et al (2015) shows a similar contribution exclusively due to the decay of 12N during irradiation of graphite. A corrigendum to Dendooven et al (2015) (Dendooven et al 2019) gives a production rate of 12N on carbon of 4.46±0.13 × 10-4 per 55 MeV proton. Therefore, the production of 13O/12N on graphite with 3He is 1.6 times higher than that of 12N on graphite with protons. Conversely, with 4He ions, the production of 12N on graphite is 2.8 times lower than 12N on graphite with protons. It is important to note that, while no production of a very short-lived nuclide is seen during irradiation of water with protons, this study shows the production of 13O/12N during irradiation of water with 3He and 4He ions with 1.3 times higher and 2.1 times smaller production rates compared to those on graphite.
For representative body tissues, such as adipose as well as for a tissue surrogate, PMMA, 13O/12N will dominate the expected PET count early-on in irradiations with 3He and 4He ions. It dominates the PET count during the first 20 s of an irradiation with 4He in adipose, muscle, and PMMA. For irradiations with 3He, the PET count dominance of 13O/12N exceeds 25 s in all materials considered. Whereas no significant dependence of the 13O/12N dominance on the carbon-oxygen ratio is observed during irradiation with 4He, the PET counts from 13O/12N dominate the total PET decays in oxygen-poor, carbon-rich materials such as adipose tissue and PMMA for at least 80 s. In oxygen-rich, carbon-poor materials like skeletal muscle, 13O/12N dominates for irradiation times up to about 25 s. This dependence of the dominance of 13O/12N on carbon-oxygen ratio for irradiation with 3He is because 15O is produced on oxygen only, in contrast to the production on both oxygen and carbon during irradiation with 4He. The independence on the carbon-oxygen ratio, observed during irradiation with 4He makes range determination using this ion beam quite robust with respect to the soft tissue type.
As the production of 13O/12N in a 3He irradiation is 3 to 4 times higher than in a 4He irradiation, from a statistical point of view, range verification using 13O/12N PET imaging will be about 2 times more accurate for a 3He irradiation compared to a 4He irradiation.
For future clinical use of the most important nuclide combination, 12N/13O, measurement of the production cross-section vs energy will be necessary for the Monte Carlo calculations of reference PET activity profiles. The reference PET activity provides
6 Summary and Outlook
131
will be used for detection of range deviations by comparing it with the measured one. One way to reduce the uncertainties in the measurement is to increase the efficiency of the detectors both for the 511 keV annihilation photons and 4.4 MeV gamma photons of 12N. Alternatively, one might consider measuring the activity yield in suitable reference materials and extrapolating such measurements to the patient anatomy. This approach is similar to the measurements required for implementation of the pencil beam dose calculation algorithm in treatment planning systems.
6.2.2 Imaging of short-lived positron emitters
The observation of a very short-lived contribution with a half-life of about 10 ms, attributed to 12N and/or 13O, during irradiation of both carbon-rich and oxygen-rich targets with helium ions as presented in chapter 3 implies that feedback on the position of the Bragg peak could be obtained faster, compared to imaging longer-lived nuclides such as 15O, by imaging this contribution. In chapter 4 of this thesis, the near real-time verification capabilities, especially the precision of 12N activity range measurements, of helium beam radiotherapy by PET imaging of this very short-lived contribution was investigated. The PET system used in the study is 1/6 of a Siemens Biograph mCT clinical scanner with custom-modified detectors. Each panel is 21 x 21 cm2 and composed of a 4 x 4 array of PMT-based block detectors. The custom modifications on the PMT base were implemented to ensure good detector performance under the high radiation levels present during the irradiation.
The PMTs were switched off during the beam-on as a protective measure against the high radiation flux during the beam-on periods. Although the PMT becomes operational within 300 µs, a recovery time of about 25 ms was observed. This recovery has been accounted for in the data analysis through a time-dependent coincidence recovery factor applied to each coincidence event. Investigations into the origin of the recovery effect seen in the data show that there is a gain shift when the PMTs are switched on after the beam pulse. Thus, the fraction of detector signals falling inside the pre-defined energy window is reduced. For future applications, a PMT gain shift correction could be implemented in either hardware or software, leading to 30% more counts and thus about 15% better precision in range measurement as the precision is largely determined by the counting statistics.
The standard deviation of the activity ranges as a metric for range precision, determined from irradiations on PMMA was found to be 9.0 and 4.1 mm (1σ) with 1.3 × 1074He ions per pulse and 6.6 × 1074He ions per pulse respectively. When considering 4.0 × 1074He ions, which is about the intensity of the most intense distal layer spot in a helium therapy plan, a range verification precision in PMMA of 5.2 mm (1σ) is derived from a power law fit to the values obtained for irradiations between 1.3 × 107 and 6.6 × 1084He ions per pulse. The experimental results highlight the prospects of obtaining 1.8 mm (1σ) precision in range verification, within 50 ms into an irradiation and when using a scanner with 29% solid angle, by in-beam PET imaging of the very short-lived positron decay contributions when summed over about 10 distal layer spots during helium therapy. For summing the same number of spots and number of protons, a much higher precision of 0.9 mm (1𝜎) can be obtained when using a scanner with 57% solid angle coverage and a beam spot duration much shorter than 10 ms.
6.3 Real-time imaging of short-lived positron emitters during
proton therapy
The most important short-lived positron emitter produced during proton therapy is 12N (T1/2 = 11 ms) (Dendooven et al 2015 and Dendooven et al 2019). The 12N decay will
6.3 Real-time imaging of short-lived positron emitters during proton therapy dominate the PET count for up to 70 s during irradiation of carbon-rich tissues with protons. A proof-of-principle study on the use of this important nuclide for near real-time proton range verification using a small PET scanner (6.5 × 6.5 cm2 modules) shows that the range can be measured with a precision of 3 mm based on the detection of 4000 12N PET counts during the delivery of 2.5 × 1010 protons over a total irradiation time of 120 s (Buitenhuis et al 2017). In chapter 5 of this thesis, the precision of range measurement for the delivery of single spots by imaging the 12N activity with a larger scanner at clinical beam fluences (108-109 protons per spot) is presented. The PET scanner used is the same as that used in the work presented in chapter 4. The issues related to the recovery of the coincidence rate after switching the PMTs on were also encountered here and taken into account in the data analysis in the same way.
The standard deviation of the activity range, determined from 30 data sets from irradiations on PMMA and graphite targets, was found to be 2.5 and 2.6 mm (1σ) for 108 protons per pulse and 0.9 and 0.8 mm (1σ) with 109 protons per pulse, respectively. Extrapolating the fit of the precision vs. number of protons to smaller values, irradiation with 4 × 107 protons per pulse, equivalent to the intensity of most spots in the distal layer of a treatment plan, gives a precision of 3.4 mm (1σ), for single spot irradiations on PMMA. Analytical extrapolations of the results from this study show that using a scanner with a solid angle coverage of 57%, with optimized detector switching and spot delivery times much smaller than the 12N half-life, a range measurement precision of 2.0 mm (1σ), 1.3 mm (1σ) and 0.5 mm (1σ) within 50 milliseconds into an irradiation with 107, 108 and 109 protons per pencil beam spot can be potentially realized. Aggregated imaging of neighbouring spots or, if possible, increasing the number of protons for a few probe beam spots will enable the realization of higher range measurement precision. Furthermore, changing to a reduced spot plan (i.e. reduction in number of spots and thus an increase in the number of protons per spot) (van de Water et al 2020) will lead to an increase in the precision of range verification. For clinical implementation, it is envisaged that range verification using 12N imaging will be for the highly weighted distal layer spots and also for those spots that have been boosted to enhance count statistics.
6.4 Future developments of optimal scanners.
The results for the performance of real-time PET-based imaging of short-lived positron emitters have been obtained with a custom-modified Siemens Biograph mCT PET scanner. Three aspects in the scanner design should be considered to improve its performance in real-time monitoring of particle therapy.
Firstly, switching mechanisms with faster gain recovery need to be investigated. Several methods for gating PMTs have been employed for pulsed counting applications. These methods include pulsing the high voltage applied to the tubes and switching the bias on the PMT electrodes (Farinelli and Malvano 1958, Abe et al 2018). It is, therefore, worthwhile to perform a systematic study on the suitability and limitations of the different switching mechanisms to identify the optimal design for in-beam 12N PET monitoring at different irradiation time structures. Implementing a fast switching and gain recovery mechanism will be particularly relevant for translation of the verification technique to short-pulse high intensity irradiation conditions such as with synchrocyclotrons (Henrotin et al 2016, Klein et al 2008, Gall 2012) and under FLASH irradiations (Favaudon et al 2014, Vozenin et al 2019). FLASH irradiation is gaining some interest in radiotherapy with electrons, photons and protons. It employs ultra-high dose rate irradiations delivered in a short time, typically less than 100 milliseconds, to reduce the radiobiological effect on normal tissue relative to tumour tissue. Thus, the differential radiobiological effect can be exploited for two applications in radiotherapy: dose escalation for treatment of radio-resistant tumours and reduction of normal tissue toxicity. Furthermore, because the
6 Summary and Outlook
133
irradiation duration is shorter than physiological processes, FLASH irradiation has the potential to minimize dose delivery uncertainties caused by intra-fractional organ motion. For proton therapy applications, the optimal beam delivery time structures and beam delivery mode required to achieve the FLASH effect are still being investigated with major proton therapy vendors already providing statements on their plans and readiness for FLASH irradiations (Varian medical system 2019, IBA 2019, Mevion 2019). An exemplary time structure may entail delivering FLASH irradiation pulses at a pulse dose rate ≥ 106 Gy/s, dose-per-pulse ≥ 1 Gy, pulse repetition frequency of 100 Hz, total dose ≥ 10 Gy and a mean dose rate ≥ 100 Gy/s (review by Wilson et al (2020)). Clearly, fast switching down to 100 Hz will be necessary to prevent damage to the PMT by the high irradiation environment. As the precision for proton range verification scales with the number of protons or number of detected 12N count, the range verification precision under FLASH irradiation can be estimated using the number of protons delivered during a probe fraction of the full dose, for e.g. the first few pulses. A typical probe fraction may consist of the fraction dose from 10 neighouring distal layer spots. For the delivery of 1 Gy to the tumour volume, distal layer spot is composed of a maximum of about 2 × 108 protons. By adoption of a line scanning technique and accumulating 12N counts over 10 such spots all delivered within an irradiation period of about 3 ms, a range measurement precision of 0.6 mm (1σ) can be realized when imaging the beam-induced 12N from these spots with an optimized scanner (fast PMT switching) having a solid angle coverage of 29% as used in the experiments of this thesis. Due to the 1 kHz pulsing of synchrocyclotrons, fast switching of the PMTs at frequencies up to at least 1 kHz is required.
Secondly, better Time-of-Flight (TOF) resolution can be implemented in a future scanner. For 12N image reconstruction with the technique described in chapter 4 and 5, the use of a sufficiently good TOF resolution would allow the exclusion of activities, i.e. longer lived ones, from outside the volume containing the 12N activity. For a 12N activity volume with a minimum diameter of 5 cm, due to the large 12N positron range, a TOF resolution of 350 ps would be sufficient. The detector technology used in this thesis has a TOF resolution of 550 ps, the best availablefor a PMT-based clinical PET system. All commercial scanners with better resolution are based on SiPMs with TOF resolutions of 200 – 350 ps. So having a scanner with SiPMs would certainly be an interesting option. The big issue at the moment for clinical implementation is the radiation hardness of the SiPMs (Diblen et al 2017). SiPMs are susceptible to radiation damage, thus limiting their operation in a high radiation environment. In the meantime, some approaches are currently being investigated to overcome these limitations (for e.g. see review by Garutti and Musienko 2019). Perhaps, a solution becomes available in the near future. Alternatively, PMT-based solutions capable of providing TOF resolution comparable to or better than that of SiPM-based detectors may be investigated. One such solution consist of a combination of L(Y)SO scintillators with multianode PMTs (Hamamatsu 2020a and 2020b) capable of achieving a TOF resolution of 200 – 300 ps. A significantly better TOF resolution of less than 100 ps can be obtained with a combination consisting of a Cherenkov radiator and microchannel plate PMT (MCP-PMT) (Ota et al 2019).
In addition to a faster switching of the PMTs and use of a good TOF resolution, a further improvement in the performance can also be expected by increasing the size of the scanner by a factor of 2 – 3. The analysis presented in section 4.4 (discussion of chapter 4) and 5.4 (discussion of chapter 5) shows that increasing the scanner size will improve the sensitivity and thus the precision of the range measurement. However, the cost effectiveness of such increments remains to be seen.
6.5 Towards clinical implementation
6.5 Towards clinical implementation
For a clinical translation of the range verification technique presented in this thesis, additional investigations need to be conducted. A next step on the road to clinical implementation is experiments on inhomogeneous structures as typically seen in clinical irradiation. Such measurements would provide information on the influence of the tissue stoichiometry and density on the activity profiles and range verification performance. A second step would involve a comparison of the activity profiles measured with the scanner and that predicted by the treatment planning system. As input to the prediction by the treatment planning system, the production cross sections as a function of energy or depth in relevant materials need to be measured. Provided that there are good results from these studies, a clinical trial with real patient treatment can then be conducted.
In light of the continuous effort to mitigate range uncertainties by various research groups, the potential contribution of in vivo verification needs to be clearly defined. Concepts, at different phases of clinical development and utilization, for reducing and mitigating range uncertainties have been investigated. These concepts include more accurate pre-treatment imaging and robust optimization of treatment plans. Some of these concepts aimed at improving the understanding of the patients’ relative stopping power model such as dual energy CT (DECT) and proton radiography/computed tomography can significantly reduce range uncertainties to about 1% leading to a reduction of range margins. Nevertheless, in vivo verification techniques will still be relevant for providing useful feedback on the accuracy of the beam delivery for quality assurance and treatment documentation purposes. Utilization in the quality assurance framework may include providing triggers for corrective strategies such as daily adaptive therapy. Following such adaptation of the treatment plan, quality assurance still needs to be provided prior to real treatment. As the patients remain on the treatment position, a fast quality assurance of the adapted treatment plan is necessary. In vivo range verification together with Monte Carlo recalculation on the patient anatomy using the log-files of the dose delivery an be used for quality assurance (Albertini et al 2019).
PET imaging of short-lived positron emitters will remain an attractive option for in vivo range verification when compared with prompt gamma detection techniques. It has the potential to achieve comparable range measurement precision especially when aggregating over multiple distal layer spots as often done with current prompt gamma detection prototypes. A better performance can be expected for PET-based imaging when transitioning to high intensity irradiation. Intense production of secondary radiation in such an irradiation scheme would lead to a saturation of the currently available prompt gamma detection prototypes. Whereas, given the delayed photon detection of PET imaging, PET-based verification techniques will be less subsceptible to such high intense radiation fluxes. The technique would also benefit from the well-established clinical imaging technology that PET offers. This implies less effort in the development of optimal technology. Considering that the basic PET technology is needed for other applications besides range verification, there will be a higher interest in the technique by equipment vendors. Prior to clinical deployment, issues related to scanner pulsing, geometrical optimization and technical integration of the scanner to the irradiation setup still need to be addressed in future studies.
6 Summary and Outlook
135
6.6 References
Albertini F, Matter M, Nenoff L, Zhang Y and Lomax A 2019 Online daily adaptive proton therapy
BJR 93 20190594
Abe Y, Nakajima N, Sakaguchi Y, Arikawa Y, Mirfayzi S R, Fujioka S, Taguchi T, Mima K, Yogo A, Nishimura H, Shiraga H and Nakai M 2018 A multichannel gated neutron detector with reduced afterpulse for low-yield neutron measurements in intense hard X-ray backgrounds Rev.
Sci. Instrum. 89 10I114
Buitenhuis, H., Diblen, F., Brzezinski, K., Brandenburg, S. and Dendooven, P. (2017). Beam-on imaging of short-lived positron emitters during proton therapy. Phys. Med. Biol., 62 4654-4672. Dendooven, P., Buitenhuis, H., Diblen, F., Heeres, P., Biegun, A., Fiedler, F., van Goethem, M.,
van der Graaf, E. and Brandenburg, S. (2015). Short-lived positron emitters in beam-on PET imaging during proton therapy. Phys. Med. Biol. 60 8923-8947.
Dendooven, P., Buitenhuis, H., Diblen, F., Heeres, P., Biegun, A., Fiedler, F., van Goethem, M., van der Graaf, E. and Brandenburg, S. (2019). Corrigendum: Short-lived positron emitters in beam-on PET imaging during proton therapy (2015 Phys. Med. Biol. 60 8923). Phys. Med. Biol.
64 129501.
Diblen F, Buitenhuis T, Solf T, Rodrigues P, van der Graaf E, van Goethem M-J, Brandenburg S and Dendooven P 2017 Radiation hardness of dsipm sensors in a proton therapy radiation environment IEEE Trans Nucl Sci 64 1891–6
Durante M and Paganetti H 2016 Nuclear physics in particle therapy: a review Rep. Prog. Phys. 79 096702
Farinelli U and Malvano R 1958 Pulsing of Photomultipliers Rev. Sci. Instrum. 29 699–701
Gall K P 2012 The single-room ion beam facility Ion Beam Therapy ed U Linz (Berlin: Springer– Verlag) pp 661–72
Garutti E and Musienko Yu 2019 Radiation damage of SiPMs Nucl Instrum Methods Phys Res A 926 69–84
Grün, R., Friedrich, T., Krämer, M., Zink, K., Durante, M., Engenhart-Cabillic, R. and Scholz, M. (2015). Assessment of potential advantages of relevant ions for particle therapy: A model based study. Med Phys 42 1037-1047.
Hamamatsu, 2020a. [online] Hamamatsu.com. Available at: <https://www.hamamatsu.com/resources/pdf/etd/H12700_H14220_TPMH1379E.pdf> [Accessed 5 June 2020].
Hamamatsu, 2020b. [online] Hamamatsu.com. Available at: <https://www.hamamatsu.com/resources/pdf/etd/R10754-07-M16_TPMH1364E.pdf > [Accessed 5 June 2020].
Henrotin, S., Forton, E., Jongen, Y., Kleeven, W., & Walle, J.V. (2016). Commissioning and testing of the first IBA S2C2 (No. TUP07). Cyclotron Technology. P. 115
IBA. (2019, June 11). Flash Irradiation Delivered In A Proteus®ONE Treatment Room [Press release]. Retrieved from https://iba-worldwide.com/content/flash-irradiation-delivered-proteus-one-treatment-room.
Jermann, M. (2017). Particle Therapy Patient Statistics (per end of 2016). [online] Ptcog.ch. Available at: https://www.ptcog.ch/archive/patient_statistics/Patientstatistics-updateDec2016.pdf
[Accessed 1 Aug. 2018].
Kempe, J., Gudowska, I. and Brahme, A. (2006). Depth absorbed dose and LET distributions of therapeutic H1, He4, Li7, and C12 beams. Med Phys 34 183-192.
Klein E, Low D and Zheng Y 2008 Design and shielding considerations for the world’s first compact synchrocyclotron proton therapy unit Med. Phys. 35 2812
Knäusl, B., Fuchs, H., Dieckmann, K. and Georg, D. (2016). Can particle beam therapy be improved using helium ions? – a planning study focusing on paediatric patients. Acta Oncol 55 751-759.
6.6 References
Krämer, M., Scifoni, E., Schuy, C., Rovituso, M., Tinganelli, W., Maier, A., Kaderka, R., Kraft-Weyrather, W., Brons, S., Tessonnier, T., Parodi, K. and Durante, M. (2016). Helium ions for radiotherapy? Physical and biological verifications of a novel treatment modality. Med Phys 43 1995-2004.
Mairani, A., Dokic, I., Magro, G., Tessonnier, T., Kamp, F., Carlson, D., Ciocca, M., Cerutti, F., Sala, P., Ferrari, A., Böhlen, T., Jäkel, O., Parodi, K., Debus, J., Abdollahi, A. and Haberer, T. (2016). Biologically optimized helium ion plans: calculation approach and its in vitro validation.
Phys. Med. Biol. 61 4283-4299.
Mevion. (2019, September 16). Focus On The Future : FLASH Therapy | Mevion Medical Systems [Press release]. Retrieved from https://www.mevion.com/newsroom/press-releases/focus-future-flash-therapy
Ota R, Nakajima K, Ogawa I, Tamagawa Y, Shimoi H, Suyama M and Hasegawa T 2019 Coincidence time resolution of 30 ps FWHM using a pair of Cherenkov-radiator-integrated MCP-PMTs Phys. Med. Biol. 64 07LT01
Tessonnier, T., Mairani, A., Chen, W., Sala, P., Cerutti, F., Ferrari, A., Haberer, T., Debus, J. and Parodi, K. (2018). Proton and helium ion radiotherapy for meningioma tumors: A Monte Carlo-based treatment planning comparison. Radiat Oncol 13 2.
Tommasino, F., Scifoni, E. and Durante, M. New Ions for Therapy. Int. J. Part. Ther. (2015) 2:428-438.
van de Water S, Belosi M F, Albertini F, Winterhalter C, Weber D C and Lomax A J 2020 Shortening delivery times for intensity-modulated proton therapy by reducing the number of proton spots: an experimental verification Phys. Med. Biol. 65 095008
Varian medical systems. (2019, April 3). Varian Discloses First Preclinical Results of Potential Major
Breakthrough in Cancer Treatment [Press release]. Retrieved from
https://www.varian.com/about-varian/newsroom/press-releases/varian-forms-flashforward-consortium-study-ultra-high-dose
Wilson JD, Hammond EM, Higgins GS and Petersson K 2020 Ultra-High Dose Rate (FLASH) Radiotherapy: Silver Bullet or Fool's Gold? Front. Oncol. 9 1563.
